PlainLink

MeatballWiki | RecentChanges | Random Page | Indices | Categories

PlainLink is an attempt to make a wiki that supports the natural 'linking' that occurs in a person's mind when they read or hear regular, plain text.

Claims

Human language has historically been ReaderDefined?. When you read a document, it is YOU, the reader that decides what each word and phrase means because of your history and exposure. Because of the influence of others, the community you keep (your friends and enemies) also effect these definitions.
CamelCase, []'s or other 'forced' linking tends to create an unnatural, WriterDefined? language.
Communities naturally define Terms as words, phrases and titles through private usage and public standards.
Word definitions are set by dictionary, thesaurus and technical reference but vary across time because of popular usage.
Phrase definitions are set by 'standards' such as books, lyrics, poems, TV, movies, news, etc.
Titles are the proper names of things.
Some Term usage is restricted by TradeMark or CopyRight law.
Humans link Terms to Definitions in their heads as they read or listen. It is almost like a LinkLanguage without technology.
Foreigners (like InterWiki) influence definitions according to community trust and acceptance.
The claim that... "PlainLink ... wants the reader to decide instead of the Author" should be considered in the reality that...
- The Author always produces a text or a presentation before it can be presented to an Audience.
  - Good Authors strive to tailor their message based on the "know your Audience" maxim in order to communicate as effectively and efficiently as possible.
  - An Author's knowledge of the Audience is generally less than perfect and is at best "averaged" out (watered down) by the distribution of the Audience's communication abilities (to Read or Listen effectively)
  - The Author inevitably must create the message before the Audience can receive it.
- The practical result is that the Author and NOT the Audience decides.
- -- HansWobbe

Technical

: Terms may contain spaces, are case insensitive, never need CamelCase or []'s or '-' or '_' or an type of LinkSyntax?. Each term links whenever the database contains a page title that matches that term. This means most page titles will contain spaces, but the href for the URL could still replace the space with a '_' to avoid the ugly %20 in the address bar.

Example

For instance, if there were a page defining the Term "Corporate Goals", then any other page that contained the phrase "corporate goals" in it's text (irrespective of case) would generate a hyperlink to the "Corporate Goals" page.

So, by writing:

: Corporate goals have erroneously been separated from governmental and communal goals.

Would cause the "Corporate Goals" page to link in the rendered HTML even though it is not surrounded by []'s, is not CamelCase, and has no _ or - between words.

By doing this, the definitions made by the community become transparent - or 'exposed' to those that are not already "in the know" about their existence.

A perfect implementation will be much more work.

We would also want to capture singular and plural variations with and without 'helper' words:

: corporate goal
: goals of incorporation
: goal of a corporation
: incorporate with the goal
: incorporate for the purpose
: goal of any corporation
: reason corporations exist
: goals of corporations
: why we incorporate
: goals of some corporations
: why groups organize as a unified body
: goals of all corporations
: ... seemingly infinite variations

Some of this can be done with regular expressions, but in the end, especially as synonyms are employed will require a parser and language analysis system maybe similar to what is used in text adventure games that try to determine what the player means.

Because of the high complexity it may seem Terms with spaces would be useless for now, but the applicability comes in those that are well known and do not vary or vary only slightly.

Implementations

http://www.mediawiki.org/wiki/Extension:AutoLink implements the first 'easy' part of PlainLink, but the word variations and suffixes are not yet addressed.

Research

http://senserelate.sourceforge.net Word sense disambiguation by measuring the semantic similarity between a word and its neighbors.

non-implementations

CleanLinking requires []'s
Oddmuse:Link_All_Words_Extension won't link page titles containing spaces.
ProWiki claims to do this, but even the example on it's homepage uses "TestPage" instead of "Test Page". Does it really work if a page title contains a space character?
GaGaParser and GakiGaki? require spaces in text be replaced by '-' or '_'.
GeboGebo at http://GeboGebo.org require spaces in text be replaced by '-' or '_'.
http://wiki.splitbrain.org/plugin:autolink?s=autolink munges the text by inserting [] chars around the Titles it finds. What a mess.

similar examples

the NY Times is doing something like this... check it out, click on random words and see what you get popping up. The challenge with this approach is that it is a hypertext free for all. A machine is determining what to link to rather than human beings intentionally making link.

I looked at http://NYTimes.com and didn't see what you describe. PlainLink doesn't want a machine to decide, it wants the readers to decide instead of the author - just as it has always been.

go to any NY Times article and pick a word and double click on it. Why does the reader get higher priority than the author? That doesn't make sense to me. This is interesting because I have often been challenged in the making of hypertext soup, but I really do believe that the author has an important task in intentional linking.

Thanks, I didn't think to double-click.

The reader should have higher priority only because the reader has *always* been the person that determines the meaning of spoken or written text. Think of reading the same newspaper on paper. Who decides what *you* think the word defense means? It is *you* that decides. This seems extremely chaotic, but it's not so bad because, although you could choose to believe defense means garbanzo beans, it would only cause you to be unable to understand anyone else, so people tend to define words and terms together - allowing others to shape the language through "common usage" etc. It is a kind of free-form collaborative agreement that people generally subscribe to because they are otherwise on their own.

The fact that readers and listeners interpret an author's intention through their own lenses is, of course, a challenge for writers who want to relay a specific message (and what writer doesnt' want this?), but I think it might be worth salvaging.

So, PlainLink is actually about CollaborativeDefinitions? because that is how it has always worked. But this makes PlainLink much more complicated than is described here because it also means every member of the community needs to be able to make a certain amount of impact on each definition without letting any one person to fully dominate.

The owners of the NY Times website are in absolute control over the possible wishes of the users/consumers, so what they have there is not quite 'right' either, though it does come close. Thanks for asking these questions, it has helped me think more about this and realize it is even larger than I imagined.